NVIDIA's GPU Technology Conference returned to San Jose from March 16-19, 2026, and Jensen Huang delivered what analysts are calling the most consequential tech keynote of the decade.
Groq 3 LPU: The $20B Chip Ships
Following NVIDIA's $20B acquisition of Groq in late 2025, the Groq 3 LPU was officially unveiled. Shipping Q3 2026, it is built specifically for inference workloads. The Groq 3 LPX rack pairs 256 LPUs alongside Vera Rubin for hybrid training and inference.
Autonomous Vehicles Take Center Stage
Five major OEMs announced adoption of NVIDIA Drive Hyperion. The Uber partnership will power robotaxi deployments across 28 cities on 4 continents by 2028.
Our CTO Hrishikesh Baidya has been tracking how Groq 3's inference cost curves could materially shift pricing for mid-market AI automation by late 2026.
Building AI-Powered Software in 2026?
Our team helps businesses design AI pipelines that take advantage of improving infrastructure economics.
Talk to Our Team
